This tutorial assumes you are running Linux and know your way around it.
Install the Java Run Time Envrionment (JRE)
For installing the JRE, please consult this webpage as it provides an easy, step-by-step process to install JRE.
To verify the JRE has been installed correctly, run:
This should output something along the line with these:
Install python libraries and dependencies
Note: we are still using python2 based dependencies. Therefore, make sure
when you run
pip, you are using python2 as your default python and install the
corresponding python2 dependencies.
- Install kafka-python
To verify the library has been installed correctly, start python shell and run:
If no error pops up, that means the library is all set.
- Install avro
To verify the library has been installed correctly, run this in a python shell:
This should not raise any error if the library is installed correctly.
Download and setup system path for Kafka
We are using an older version of Apache Kafka. It can be downloaded from this link.
After download is complete, extract the content into a directory of your
choice. The directory will have
bin that has all the executables and
that contains all the configuration files for the executables.
We need to add the needed executables into system path. To do so, add
this line to the end of your
Save and exit and do:
Then this path will be added to your system path ONLY for this terminal session. To make this permanent, you have to log back out and in again.
Make sure you have these executables in your path:
Download custom scripts and schemas
First off, download the custom scripts from here and extract the contents. Let's go through the contents one by one.
This script starts off an instance of zookeeper and a Kafka broker. There are several variables to specify for it to run correctly. The specifications will be discussed in the Setup a zookeeper and a Kafka broker section.
You can use this script to stop the zookeeper and the broker instance you created.
These two scripts use the
python-kafka and the
avro library you just
installed. One consumes the CAN messages and the other consumes the gps
messages from the broker you set up. The details on how to use these two scripts
will be discussed in later sections.
This directory contains the avro schemas we defined for different message types.
Have your data ready
You can transfer the
kafka-logs directory from your ISOBlue 2.0 device to a
local directory of your choice. Alternatively, you can download some
from the Public Data section from this site to test.
Setup a zookeeper and a Kafka broker
For this part, you will need the
There are two parameters you need to change in this script.
This directory specifies the configuration files directory for the zookeeper and the broker. It usually should be changed to a absolute path of where you put you extracted the Kafka directory.
This directory should contain the absolute path that point to whereever you
After these two parameters are set, run this in your terminal:
If no error pop up, then an instance of a zookeeper and a broker should be up.
Be default, the Kafka broker will roll over the current loaded logs after a retention period of 7 days (168 hours). There are various retention configuration you can set to tailor your retention needs. For example, if you set
log.retention.hours=1to have Kafka delete your current logs and start over fresh everytime an hour has passed. For different parameters for changing the default retention policy, please consult the Kafka broker configs.
Consume Kafka messages
Once you have the zookeeper and broker up, we can start consuming messages from
them. Make sure the
schema directory is in the same directory as where
the scripts reside.
Consume CAN messages
To consume the CAN messages, you will use the
You will need to specify a
topic and an
auto_offset_reset parameter. The
topics for consuming CAN messages are:
And the available
auto_offset_reset parameters are:
For consuming the CAN messages from the start of the tractor bus log, you will run:
As the script defaults to consume from the beginning of the log from a given topic.
The script will then output the consumed logs that has timestamp, pgn, and the data payload bytes and these fields are space-separated.
You can use
> to redirect the script output to a file, i.e.:
The script will exit once it consumes all the messages.
Consume GPS messages
kafka_gps_consumer.py is not written as rigorous as
kafka_can_consumer.py. The topic is set to
gps and the
parameter is already set to
earliest in the script. To consume the GPS
messages and store them in a file, just run:
Stop zookeeper and broker
After you think you are done using the zookeeper and broker, you can start them by running the custom script: